tira-asr / README.md
markjosims's picture
Upload dataset
6c39431 verified
metadata
dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: start
      dtype: float64
    - name: end
      dtype: float64
    - name: transcription
      dtype: string
    - name: eaf_source
      dtype: string
    - name: wav_source
      dtype: string
    - name: raw_transcription
      dtype: string
    - name: clip
      dtype: string
    - name: wav_rawpath
      dtype: string
    - name: path
      dtype: string
    - name: allosaurus
      dtype: string
    - name: clap_ipa_cos_sim
      dtype: float64
    - name: wada_snr
      dtype: float64
    - name: nist_stnr
      dtype: float64
    - name: speaker-diarization-3.1
      dtype: string
    - name: voice-activity-detection
      dtype: string
    - name: whisper-large-v3
      dtype: string
    - name: clapipa-transcription-allosaurus
      dtype: float64
    - name: vad_s
      dtype: float64
    - name: drz_s
      dtype: float64
    - name: trans_len
      dtype: int64
    - name: pcnt_speech
      dtype: float64
    - name: trans_len_sq
      dtype: float64
    - name: trans_len_log
      dtype: float64
    - name: cos_sim_softmax
      dtype: float64
    - name: cos_sim_log
      dtype: float64
    - name: duration
      dtype: float64
  splits:
    - name: train
      num_bytes: 1106517483.8
      num_examples: 16384
    - name: validation
      num_bytes: 131592822.088
      num_examples: 2048
    - name: test
      num_bytes: 76685401.888
      num_examples: 2048
  download_size: 1290401128
  dataset_size: 1314795707.776
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*